Some notes on AI and “fine art” imagery

A couple of weeks ago I had a post entitled “Friday Fotos: The Last Frontier of AI.” I was interested in whether or not a certain approach I’d been using to create images with ChatGPT could produce “fine art” images, as opposed to illustrations or popular art of various kinds. The particular images I developed for that post (there were five), while interesting, were not particularly compelling. So I went on to display eleven other images I’d created with ChatGPT, including some of the images which that had motivated the post in the first place. I then ranked the images among themselves and decided that the images I’d created specifically for the post ranked near the bottom.

So, while the approach that motivated that post cannot be called a success, the post as a whole has raised the question: Can ChatGPT (be used to) create “fine art” images? I put “fine art” in quotes because the term itself is problematic. It’s not as though there are identifiable characteristics such that any image exhibiting them is a fine art image. The notion of fine art as opposed to folk art or popular art or (mere) illustrations is a cultural convention, one that Marcel Duhamps exploded in 1917 when entered a urinal into the inaugural exhibition of the Society of Independent Artists. He called it Fountain and attributed it to “R. Mutt.” Fine art is simply the art society has decided deserves to be treated in a certain way, no more, no less. If you decided that a common urinal should be treated in that way, then it becomes fine art.

Duchamp’s move was controversial, and that controversy has been reverberating ever since. I have no intention of reviewing and rehashing it here. Rather, I simply want to present a collection of images I’ve made with ChatGPT and view them with that issue reverberating in the background.

This image is one of my favorites among those I’ve created with ChatGPT:

I created it for illustrative purposes, to go on the cover of a working paper about Joseph Conrad’s Heart of Darkness, but I think the image stands on its own. If you’re familiar with the book, then resonance is obvious. It tells about a voyage up the Congo River. As for the superimposed image of the Buddha, here’s first sentence of the last paragraph: “Marlow ceased, and sat apart, indistinct and silent, in the pose of a meditating Buddha.”

I should note, and this is important, I had ChatGPT create that image from within a chat devoted to that working paper, making the entire chat (up to that point) the context in which ChatGPT created the image. I have reason to believe that it wasn’t working simply from prompt that generated that image. For a discussion of that, see this post: High-level “vibe” – Creating Imaginary Bank Notes with ChatGPT: AI as cultural technology and collective creativity.

Here’s a somewhat different image that I also like very much. It’s almost, but not completely, abstract:

The book is obvious. The rest of it? But that’s not the first image ChatGPT offered to me. This came before (and there were others before this):

If it’s fine art we’re interested in, the black and white image seems (vastly) superior to me.

Here’s an utterly different image:

I don’t remember what prompt I used to create that. But I like the image, absurd as it is, a lot. THAT’s why I like it. It’s ridiculous, but fun. Fine art? Ask me if I care. 

What about this? 

It’s one in a series of images that started out as a birthday card for a friend who is the daughter of immigrants from Sri Lanka. She’s also trained in Indian classical dancer. So the series started out with the image of a young woman dancing a solo in a Bollywood musical. Then I thought: “Why not put her in black leathers riding a Ducati?” ChatGPT had no problem with that. If we can do that, why not have her play Mary Magdelene at the cross? I had it do a number of those in various styles, Rembrandt, Salvador Dali, Picasso, and Jack Kirby. But this is my favorite image in the series:

Talk about surreal, what about this image, based on the third stanza of Andrew Marvell’s “To His Coy Mistress”?

The next two are based on photographs. The first is based on a photo of a wooded area within Jersey City. I asked ChatGPT to imagine a scene in an Indiana Jones film where Jones has just seen the first signs of a lost civilization:

“Yeah, I know, more pop culture than fine art. What can I say?” The next image is derived from a photograph of a café in a park in Hoboken as it might be in 2125:  

Coming back to earth, ChatGPT entitled this one, “Celestial dawn over a medieval city.” This is one of a bunch of images I had ChatGPT in connection with an imagined conversation between St. Augustine, Poe Leo XIV, and Kurt Gödel:

That might almost be a naturalistic image if it weren’t for that jazz in the sky. Now let’s add our three interlocutors:

Do we need to know who those guys are? Why or why not? Do we need to know what they’re talking about? What would those images look like twelve feet wide and eight feet high?

Finally, we have this:

ChatGPT created it to accompany a poem by my sister, “Everywhere is the touch.” I uploaded the poem to ChatGPT and prompted it to create an appropriate image. I like it a lot. But it seems a bit sentimental and the girl is too much like a greeting-card image. ChatGPT did that back in May of 2025. 

“I wonder what it would do now?” That’s easy to find out. I repeated the exercise earlier today. I didn’t have any particular expectation, but I suppose that, by default, I was looking like for another version of that image from last year. ChatGPT surprised me:

That’s a completely different image. It reads like an allegory of family dinner during America in the Cold Way. “I wonder what that image would look like when rendered as a pen-and-ink drawing? Voilà!

* * * * *

I’m not sure what we can (safely) conclude from this exercise. Nor am I particularly anxious about it. I could easily have added a dozen or two dozen more images to the discussion. I haven’t explored abstract expressionist imagery, but I don’t see why ChatGPT could deal with it.

I will say that I’m not particularly interested in what kinds of images ChatGPT could create “all on its own.” I’m not sure what that would mean. I suppose I could simply prompt it: “Create an image.” If it were to ask, “What kind of image,” I’d reply, “You choose.” Would it then go ahead? I suppose so, but I don’t really know, nor am I very motivated to find out.

I will note in closing, though, that editing images can be tricky. If ChatGPT had the kind of symbolic capacity that Gary Marcus keeps calling for that would improve. But the problem would always be there simply because of the mismatch between language and image. On the whole, I think the limitations are set more by the imagination, skill, and patience of the prompter than by ChatGPT.